| Line | Branch | Exec | Source |
|---|---|---|---|
| 1 | /* | ||
| 2 | * Copyright (C) 2016 foo86 | ||
| 3 | * | ||
| 4 | * This file is part of FFmpeg. | ||
| 5 | * | ||
| 6 | * FFmpeg is free software; you can redistribute it and/or | ||
| 7 | * modify it under the terms of the GNU Lesser General Public | ||
| 8 | * License as published by the Free Software Foundation; either | ||
| 9 | * version 2.1 of the License, or (at your option) any later version. | ||
| 10 | * | ||
| 11 | * FFmpeg is distributed in the hope that it will be useful, | ||
| 12 | * but WITHOUT ANY WARRANTY; without even the implied warranty of | ||
| 13 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | ||
| 14 | * Lesser General Public License for more details. | ||
| 15 | * | ||
| 16 | * You should have received a copy of the GNU Lesser General Public | ||
| 17 | * License along with FFmpeg; if not, write to the Free Software | ||
| 18 | * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA | ||
| 19 | */ | ||
| 20 | |||
| 21 | #ifndef AVCODEC_DCA_CORE_H | ||
| 22 | #define AVCODEC_DCA_CORE_H | ||
| 23 | |||
| 24 | #include "libavutil/float_dsp.h" | ||
| 25 | #include "libavutil/fixed_dsp.h" | ||
| 26 | #include "libavutil/mem_internal.h" | ||
| 27 | #include "libavutil/tx.h" | ||
| 28 | |||
| 29 | #include "avcodec.h" | ||
| 30 | #include "get_bits.h" | ||
| 31 | #include "dca.h" | ||
| 32 | #include "dca_exss.h" | ||
| 33 | #include "dcadsp.h" | ||
| 34 | #include "dcadct.h" | ||
| 35 | #include "dcamath.h" | ||
| 36 | #include "dcahuff.h" | ||
| 37 | #include "synth_filter.h" | ||
| 38 | |||
| 39 | #define DCA_CHANNELS 7 | ||
| 40 | #define DCA_SUBBANDS 32 | ||
| 41 | #define DCA_SUBBANDS_X96 64 | ||
| 42 | #define DCA_SUBFRAMES 16 | ||
| 43 | #define DCA_SUBBAND_SAMPLES 8 | ||
| 44 | #define DCA_PCMBLOCK_SAMPLES 32 | ||
| 45 | #define DCA_LFE_HISTORY 8 | ||
| 46 | #define DCA_ABITS_MAX 26 | ||
| 47 | |||
| 48 | #define DCA_CORE_CHANNELS_MAX 6 | ||
| 49 | #define DCA_DMIX_CHANNELS_MAX 4 | ||
| 50 | #define DCA_XXCH_CHANNELS_MAX 2 | ||
| 51 | #define DCA_EXSS_CHANNELS_MAX 8 | ||
| 52 | #define DCA_EXSS_CHSETS_MAX 4 | ||
| 53 | |||
| 54 | #define DCA_FILTER_MODE_X96 0x01 | ||
| 55 | #define DCA_FILTER_MODE_FIXED 0x02 | ||
| 56 | |||
| 57 | enum DCACoreAudioMode { | ||
| 58 | DCA_AMODE_MONO, // Mode 0: A (mono) | ||
| 59 | DCA_AMODE_MONO_DUAL, // Mode 1: A + B (dual mono) | ||
| 60 | DCA_AMODE_STEREO, // Mode 2: L + R (stereo) | ||
| 61 | DCA_AMODE_STEREO_SUMDIFF, // Mode 3: (L+R) + (L-R) (sum-diff) | ||
| 62 | DCA_AMODE_STEREO_TOTAL, // Mode 4: LT + RT (left and right total) | ||
| 63 | DCA_AMODE_3F, // Mode 5: C + L + R | ||
| 64 | DCA_AMODE_2F1R, // Mode 6: L + R + S | ||
| 65 | DCA_AMODE_3F1R, // Mode 7: C + L + R + S | ||
| 66 | DCA_AMODE_2F2R, // Mode 8: L + R + SL + SR | ||
| 67 | DCA_AMODE_3F2R, // Mode 9: C + L + R + SL + SR | ||
| 68 | |||
| 69 | DCA_AMODE_COUNT | ||
| 70 | }; | ||
| 71 | |||
| 72 | enum DCACoreExtAudioType { | ||
| 73 | DCA_EXT_AUDIO_XCH = 0, | ||
| 74 | DCA_EXT_AUDIO_X96 = 2, | ||
| 75 | DCA_EXT_AUDIO_XXCH = 6 | ||
| 76 | }; | ||
| 77 | |||
| 78 | enum DCACoreLFEFlag { | ||
| 79 | DCA_LFE_FLAG_NONE, | ||
| 80 | DCA_LFE_FLAG_128, | ||
| 81 | DCA_LFE_FLAG_64, | ||
| 82 | DCA_LFE_FLAG_INVALID | ||
| 83 | }; | ||
| 84 | |||
| 85 | typedef struct DCADSPData { | ||
| 86 | union { | ||
| 87 | struct { | ||
| 88 | DECLARE_ALIGNED(32, float, hist1)[1024]; | ||
| 89 | DECLARE_ALIGNED(32, float, hist2)[64]; | ||
| 90 | } flt; | ||
| 91 | struct { | ||
| 92 | DECLARE_ALIGNED(32, int32_t, hist1)[1024]; | ||
| 93 | DECLARE_ALIGNED(32, int32_t, hist2)[64]; | ||
| 94 | } fix; | ||
| 95 | } u; | ||
| 96 | int offset; | ||
| 97 | } DCADSPData; | ||
| 98 | |||
| 99 | typedef struct DCACoreDecoder { | ||
| 100 | AVCodecContext *avctx; | ||
| 101 | GetBitContext gb; | ||
| 102 | GetBitContext gb_in; | ||
| 103 | |||
| 104 | // Bit stream header | ||
| 105 | int crc_present; ///< CRC present flag | ||
| 106 | int npcmblocks; ///< Number of PCM sample blocks | ||
| 107 | int frame_size; ///< Primary frame byte size | ||
| 108 | int audio_mode; ///< Audio channel arrangement | ||
| 109 | int sample_rate; ///< Core audio sampling frequency | ||
| 110 | int bit_rate; ///< Transmission bit rate | ||
| 111 | int drc_present; ///< Embedded dynamic range flag | ||
| 112 | int ts_present; ///< Embedded time stamp flag | ||
| 113 | int aux_present; ///< Auxiliary data flag | ||
| 114 | int ext_audio_type; ///< Extension audio descriptor flag | ||
| 115 | int ext_audio_present; ///< Extended coding flag | ||
| 116 | int sync_ssf; ///< Audio sync word insertion flag | ||
| 117 | int lfe_present; ///< Low frequency effects flag | ||
| 118 | int predictor_history; ///< Predictor history flag switch | ||
| 119 | int filter_perfect; ///< Multirate interpolator switch | ||
| 120 | int source_pcm_res; ///< Source PCM resolution | ||
| 121 | int es_format; ///< Extended surround (ES) mastering flag | ||
| 122 | int sumdiff_front; ///< Front sum/difference flag | ||
| 123 | int sumdiff_surround; ///< Surround sum/difference flag | ||
| 124 | |||
| 125 | // Primary audio coding header | ||
| 126 | int nsubframes; ///< Number of subframes | ||
| 127 | int nchannels; ///< Number of primary audio channels (incl. extension channels) | ||
| 128 | int ch_mask; ///< Speaker layout mask (incl. LFE and extension channels) | ||
| 129 | int8_t nsubbands[DCA_CHANNELS]; ///< Subband activity count | ||
| 130 | int8_t subband_vq_start[DCA_CHANNELS]; ///< High frequency VQ start subband | ||
| 131 | int8_t joint_intensity_index[DCA_CHANNELS]; ///< Joint intensity coding index | ||
| 132 | int8_t transition_mode_sel[DCA_CHANNELS]; ///< Transient mode code book | ||
| 133 | int8_t scale_factor_sel[DCA_CHANNELS]; ///< Scale factor code book | ||
| 134 | int8_t bit_allocation_sel[DCA_CHANNELS]; ///< Bit allocation quantizer select | ||
| 135 | int8_t quant_index_sel[DCA_CHANNELS][DCA_CODE_BOOKS]; ///< Quantization index codebook select | ||
| 136 | int32_t scale_factor_adj[DCA_CHANNELS][DCA_CODE_BOOKS]; ///< Scale factor adjustment | ||
| 137 | |||
| 138 | // Primary audio coding side information | ||
| 139 | int8_t nsubsubframes[DCA_SUBFRAMES]; ///< Subsubframe count for each subframe | ||
| 140 | int8_t prediction_mode[DCA_CHANNELS][DCA_SUBBANDS_X96]; ///< Prediction mode | ||
| 141 | int16_t prediction_vq_index[DCA_CHANNELS][DCA_SUBBANDS_X96]; ///< Prediction coefficients VQ address | ||
| 142 | int8_t bit_allocation[DCA_CHANNELS][DCA_SUBBANDS_X96]; ///< Bit allocation index | ||
| 143 | int8_t transition_mode[DCA_SUBFRAMES][DCA_CHANNELS][DCA_SUBBANDS]; ///< Transition mode | ||
| 144 | int32_t scale_factors[DCA_CHANNELS][DCA_SUBBANDS][2]; ///< Scale factors (2x for transients and X96) | ||
| 145 | int8_t joint_scale_sel[DCA_CHANNELS]; ///< Joint subband codebook select | ||
| 146 | int32_t joint_scale_factors[DCA_CHANNELS][DCA_SUBBANDS_X96]; ///< Scale factors for joint subband coding | ||
| 147 | |||
| 148 | // Auxiliary data | ||
| 149 | int prim_dmix_embedded; ///< Auxiliary dynamic downmix flag | ||
| 150 | int prim_dmix_type; ///< Auxiliary primary channel downmix type | ||
| 151 | int prim_dmix_coeff[DCA_DMIX_CHANNELS_MAX * DCA_CORE_CHANNELS_MAX]; ///< Dynamic downmix code coefficients | ||
| 152 | |||
| 153 | // Core extensions | ||
| 154 | int ext_audio_mask; ///< Bit mask of fully decoded core extensions | ||
| 155 | |||
| 156 | // XCH extension data | ||
| 157 | int xch_pos; ///< Bit position of XCH frame in core substream | ||
| 158 | |||
| 159 | // XXCH extension data | ||
| 160 | int xxch_crc_present; ///< CRC presence flag for XXCH channel set header | ||
| 161 | int xxch_mask_nbits; ///< Number of bits for loudspeaker mask | ||
| 162 | int xxch_core_mask; ///< Core loudspeaker activity mask | ||
| 163 | int xxch_spkr_mask; ///< Loudspeaker layout mask | ||
| 164 | int xxch_dmix_embedded; ///< Downmix already performed by encoder | ||
| 165 | int xxch_dmix_scale_inv; ///< Downmix scale factor | ||
| 166 | int xxch_dmix_mask[DCA_XXCH_CHANNELS_MAX]; ///< Downmix channel mapping mask | ||
| 167 | int xxch_dmix_coeff[DCA_XXCH_CHANNELS_MAX * DCA_CORE_CHANNELS_MAX]; ///< Downmix coefficients | ||
| 168 | int xxch_pos; ///< Bit position of XXCH frame in core substream | ||
| 169 | |||
| 170 | // X96 extension data | ||
| 171 | int x96_rev_no; ///< X96 revision number | ||
| 172 | int x96_crc_present; ///< CRC presence flag for X96 channel set header | ||
| 173 | int x96_nchannels; ///< Number of primary channels in X96 extension | ||
| 174 | int x96_high_res; ///< X96 high resolution flag | ||
| 175 | int x96_subband_start; ///< First encoded subband in X96 extension | ||
| 176 | int x96_rand; ///< Random seed for generating samples for unallocated X96 subbands | ||
| 177 | int x96_pos; ///< Bit position of X96 frame in core substream | ||
| 178 | |||
| 179 | // Sample buffers | ||
| 180 | unsigned int x96_subband_size; | ||
| 181 | int32_t *x96_subband_buffer; ///< X96 subband sample buffer base | ||
| 182 | int32_t *x96_subband_samples[DCA_CHANNELS][DCA_SUBBANDS_X96]; ///< X96 subband samples | ||
| 183 | |||
| 184 | unsigned int subband_size; | ||
| 185 | int32_t *subband_buffer; ///< Subband sample buffer base | ||
| 186 | int32_t *subband_samples[DCA_CHANNELS][DCA_SUBBANDS]; ///< Subband samples | ||
| 187 | int32_t *lfe_samples; ///< Decimated LFE samples | ||
| 188 | |||
| 189 | // DSP contexts | ||
| 190 | DCADSPData dcadsp_data[DCA_CHANNELS]; ///< FIR history buffers | ||
| 191 | DCADSPContext *dcadsp; | ||
| 192 | DCADCTContext dcadct; | ||
| 193 | AVTXContext *imdct[2]; | ||
| 194 | av_tx_fn imdct_fn[2]; | ||
| 195 | SynthFilterContext synth; | ||
| 196 | AVFloatDSPContext *float_dsp; | ||
| 197 | AVFixedDSPContext *fixed_dsp; | ||
| 198 | |||
| 199 | // PCM output data | ||
| 200 | unsigned int output_size; | ||
| 201 | void *output_buffer; ///< PCM output buffer base | ||
| 202 | int32_t *output_samples[DCA_SPEAKER_COUNT]; ///< PCM output for fixed point mode | ||
| 203 | int32_t output_history_lfe_fixed; ///< LFE PCM history for X96 filter | ||
| 204 | float output_history_lfe_float; ///< LFE PCM history for X96 filter | ||
| 205 | |||
| 206 | int ch_remap[DCA_SPEAKER_COUNT]; ///< Channel to speaker map | ||
| 207 | int request_mask; ///< Requested channel layout (for stereo downmix) | ||
| 208 | |||
| 209 | int npcmsamples; ///< Number of PCM samples per channel | ||
| 210 | int output_rate; ///< Output sample rate (1x or 2x header rate) | ||
| 211 | |||
| 212 | int filter_mode; ///< Previous filtering mode for detecting changes | ||
| 213 | } DCACoreDecoder; | ||
| 214 | |||
| 215 | 7775 | static inline int ff_dca_core_map_spkr(DCACoreDecoder *core, int spkr) | |
| 216 | { | ||
| 217 |
2/2✓ Branch 0 taken 7655 times.
✓ Branch 1 taken 120 times.
|
7775 | if (core->ch_mask & (1U << spkr)) |
| 218 | 7655 | return spkr; | |
| 219 |
3/4✓ Branch 0 taken 48 times.
✓ Branch 1 taken 72 times.
✓ Branch 2 taken 48 times.
✗ Branch 3 not taken.
|
120 | if (spkr == DCA_SPEAKER_Lss && (core->ch_mask & DCA_SPEAKER_MASK_Ls)) |
| 220 | 48 | return DCA_SPEAKER_Ls; | |
| 221 |
3/4✓ Branch 0 taken 48 times.
✓ Branch 1 taken 24 times.
✓ Branch 2 taken 48 times.
✗ Branch 3 not taken.
|
72 | if (spkr == DCA_SPEAKER_Rss && (core->ch_mask & DCA_SPEAKER_MASK_Rs)) |
| 222 | 48 | return DCA_SPEAKER_Rs; | |
| 223 | 24 | return -1; | |
| 224 | } | ||
| 225 | |||
| 226 | 681852 | static inline void ff_dca_core_dequantize(int32_t *output, const int32_t *input, | |
| 227 | int32_t step_size, int32_t scale, int residual, int len) | ||
| 228 | { | ||
| 229 | // Account for quantizer step size | ||
| 230 | 681852 | int64_t step_scale = (int64_t)step_size * scale; | |
| 231 | 681852 | int n, shift = 0; | |
| 232 | |||
| 233 | // Limit scale factor resolution to 22 bits | ||
| 234 |
2/2✓ Branch 0 taken 471989 times.
✓ Branch 1 taken 209863 times.
|
681852 | if (step_scale > (1 << 23)) { |
| 235 | 471989 | shift = av_log2(step_scale >> 23) + 1; | |
| 236 | 471989 | step_scale >>= shift; | |
| 237 | } | ||
| 238 | |||
| 239 | // Scale the samples | ||
| 240 |
2/2✓ Branch 0 taken 4662 times.
✓ Branch 1 taken 677190 times.
|
681852 | if (residual) { |
| 241 |
2/2✓ Branch 0 taken 37296 times.
✓ Branch 1 taken 4662 times.
|
41958 | for (n = 0; n < len; n++) |
| 242 | 37296 | output[n] += clip23(norm__(input[n] * step_scale, 22 - shift)); | |
| 243 | } else { | ||
| 244 |
2/2✓ Branch 0 taken 5152816 times.
✓ Branch 1 taken 677190 times.
|
5830006 | for (n = 0; n < len; n++) |
| 245 | 5152816 | output[n] = clip23(norm__(input[n] * step_scale, 22 - shift)); | |
| 246 | } | ||
| 247 | 681852 | } | |
| 248 | |||
| 249 | int ff_dca_core_parse(DCACoreDecoder *s, const uint8_t *data, int size); | ||
| 250 | int ff_dca_core_parse_exss(DCACoreDecoder *s, const uint8_t *data, DCAExssAsset *asset); | ||
| 251 | int ff_dca_core_filter_fixed(DCACoreDecoder *s, int x96_synth); | ||
| 252 | int ff_dca_core_filter_frame(DCACoreDecoder *s, AVFrame *frame); | ||
| 253 | av_cold void ff_dca_core_flush(DCACoreDecoder *s); | ||
| 254 | av_cold int ff_dca_core_init(DCACoreDecoder *s); | ||
| 255 | av_cold void ff_dca_core_close(DCACoreDecoder *s); | ||
| 256 | |||
| 257 | #endif | ||
| 258 |